The myotonic dystrophies (DM1 and DM2), which are the most common form of adult-onset muscular dystrophy, are autosomal dominant diseases with similar clinical presentations. Remarkably, DM1 and DM2 are caused by unstable microsatellite expansions in the untranslated regions of two different genes, DMPK and ZNF9. To explain how these non-coding expansion mutations lead to dominantly inherited neuromuscular disorders, we have proposed a toxic RNA model for the myotonic dystrophies. Transcription of the mutant DM1 (CTG)n and DM2 (CCTG)n alleles leads to the production of unusual RNA transcripts with (CUG)n and (CCUG)n repeat expansions. These expansions fold into stable double-stranded (ds) RNA structures that recruit and then sequester a family of dsRNA-binding factors, the muscleblind proteins. Because this toxic RNA model suggests that DM1 and DM2 diseases are due to loss of muscleblind protein function, we have derived muscleblind 1 (Mbnl1) knockout mice. This proposal is designed to test our working hypothesis that Mbnl1-/- knockout mice will be a useful model to examine underlying molecular mechanisms involved in myotonic dystrophy disease pathogenesis. First, we will characterize the Mbnl1-/-muscle phenotype and test the hypothesis that Mbnl1 is required for proper alternative splicing and function of the chloride channel CIC-1. Deficiency of this ion channel has been recently implicated as the cause of DM1- and DM2- associated myotonia. The stoichiometric relationship between toxic RNA and binding protein will be examined by breeding Mbnll knockout mice with lines of transgenic mice that express (CUG)n RNA at different levels. Second, the possibility that muscleblind proteins influence CIC-1 chloride channel levels by interacting with alternative splicing, and/or other, factors will be examined. Third, the hypothesis that the myotonia phenotype can berescued using recombinant adeno-associated virus mediated expression of wild type adult CIC-1 will be tested. Finally, we will investigate if additional disease-associated phenotypes result from deletion of the entire Mbnl1 gene, from tissue-specific Mbnl1 expression or from combinatorial loss of all three muscleblind (Mbnl1, Mbnl2/Mbnll, Mbnl3/Mblx) genes.